Building Real Time Edge Machine Learning Systems for High Data Rate Acquisition
Bibliographic record
Abstract
Over the past decade, a developments in radiation and photonic detectors has significantly improved their resolution, pixel density, sensitivity, and sampling rate. The increase in sampling rate corresponds to a considerable increase in generated data, the movement and storage of which requires a large number of storage units with very high bandwidth interconnects to the sensors themselves. The paradigm of edge computing, however, proposes to move the data processing closer to the source, the edge, rather than moving data to processing. Still, the computation resources are limited at the edge and it is necessary to use lean and robust algorithms. Machine learning (ML) is commonly used to identify patterns and relationships in minimally processed data. EdgeML is a combination of ML and edge computing that leverages a combination of benefits from ML algorithms and edge computing. In this paper, we demonstrate a high-speed and configurable ML model in a fully customizable EdgeML flow. Our demonstration focuses on an angular streaking detector developed for the LCLS-II project known as the CookieBox. The flow starts by emulating the CookieBox, digitizing the signals, and passing them to an optimized ML model on the FPGA. By using our ML implementation in this flow, we are able to achieve a 2.7 μs of inference latency. This flow can also be configured for other instrumentation applications that require low-latency solutions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.007 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".